Intelligent logistics sorting system and sorting method

Abstract

The invention relates to the technical field of logistics, and discloses an intelligent logistics sorting system and method, and the system comprises a goods conveying unit, an intelligent recognition unit, a central control unit, a sorting execution unit, and an equipment scheduling unit. The goods conveying unit is composed of a plurality of conveying belts which are connected with one another. The central control unit receives the cargo information of the intelligent identification unit, performs data analysis processing in combination with a preset sorting rule of the system and a current sorting task, generates a sorting instruction, sends the sorting instruction to the sorting execution unit and the equipment scheduling unit, and monitors the running state of equipment in the system in real time. And the sorting execution unit comprises a plurality of sorting mechanisms, and carries out sorting operation on the goods according to the sorting instruction of the central control unit. And the equipment scheduling unit is used for uniformly scheduling and managing the running path of the goods conveying unit, the working sequence of the sorting execution unit and the working time of the intelligent identification unit according to an instruction of the central control unit.

Description

Technical Field

[0001] The present invention relates to the field of logistics technology, and in particular to an intelligent logistics sorting system and a sorting method. Background Art

[0002] With the rapid development of the e-commerce industry, the number of logistics packages has exploded, and traditional logistics sorting systems have been unable to meet the growing sorting needs. At present, the traditional logistics sorting system has the following main problems: First, relying on manual sorting is not only inefficient but also prone to sorting errors, making it difficult to ensure the accuracy of the sorting of goods; second, most existing automated sorting equipment adopts a fixed sorting process, lacks real-time perception and intelligent analysis capabilities of goods information, and cannot flexibly adapt to goods of different sizes, weights, and shapes as well as diverse sorting needs; third, the information exchange between the various devices in the traditional sorting system is not smooth, and the collaborative operation ability is poor, resulting in low overall sorting efficiency. In addition, there is a lack of effective emergency response mechanisms when equipment fails, which easily causes interruptions in sorting operations and affects the timeliness of logistics. Therefore, there is an urgent need to design a logistics sorting system that can achieve efficient, accurate, and intelligent sorting.

[0003] The present invention aims to solve the technical problem of how to navigate a transport vehicle to its shipping location to return the package and automatically sort and pack the package when a package is detected to be severely damaged, thereby providing an intelligent logistics sorting system. Summary of the Invention

[0004] The technical solution adopted by the present invention to solve the technical problem is: an intelligent logistics sorting system, comprising: The cargo conveying unit consists of multiple interconnected conveyor belts equipped with speed regulators and photoelectric sensors to transport cargo to the sorting area. The speed regulator adjusts the conveyor belt speed in real time according to the cargo type and sorting requirements, and the photoelectric sensor detects the cargo position and passing status. The intelligent identification unit includes a visual recognition module and an RFID recognition module. The visual recognition module identifies the size, shape, color and label information of the goods, and the RFID recognition module reads the RFID tag information of the goods through wireless radio frequency; The central control unit receives cargo information from the intelligent identification unit, analyzes and processes the data based on the system's preset sorting rules and current sorting tasks, generates sorting instructions, and sends them to the sorting execution unit and equipment scheduling unit. It also monitors the operating status of the equipment in the system in real time. The sorting execution unit includes multiple sorting mechanisms, which sort the goods according to the sorting instructions of the central control unit. The sorting mechanism includes a driving device and a positioning device; The equipment scheduling unit, based on the instructions of the central control unit, uniformly schedules and manages the running path of the cargo conveying unit, the working sequence of the sorting execution unit, and the working time of the intelligent identification unit.

[0005] Optionally, the visual recognition module includes: a multispectral imaging system configured with an RGB camera and a near-infrared camera for simultaneously capturing visible light images and near-infrared images of the goods, wherein the near-infrared camera operates in a wavelength range of 700-1000nm; Image preprocessing module, which uses histogram equalization to reduce noise and enhance the original image; The deep learning recognition engine builds an object detection model based on the Faster R-CNN architecture, combines it with the ResNet-50 backbone network to extract image features, and uses a feature pyramid network to achieve multi-scale detection, identifying cargo outlines, barcodes, and text. The 3D reconstruction unit uses a structured light 3D scanner to obtain 3D point cloud data of the goods, uses an iterative closest point algorithm to perform point cloud registration, and calculates the precise dimensional parameters of the goods; The tag anti-collision processing module is based on the dynamic frame time slot ALOHA algorithm and can identify more than 80 tags at the same time.

[0006] Optionally, the driving device includes a servo motor driving system and a hydraulic driving system. The servo motor driving system uses an encoder to provide real-time feedback of motor speed and position information to achieve high-precision speed and position control. The hydraulic driving system uses a hydraulic pump to provide stable pressure and accurately adjusts the hydraulic oil flow and pressure through a hydraulic valve group to drive the executive components of the sorting mechanism; the positioning device includes a laser positioning sensor and a grating scale positioning device. The laser positioning sensor accurately measures the position of the goods and the position of the moving parts of the sorting mechanism by emitting a laser beam and receiving reflected light. The grating scale positioning device reads the scale information on the grating scale.

[0007] Optionally, the equipment scheduling unit includes: a path planning module, which generates an optimal delivery path based on the cargo destination and current equipment status using an improved shortest path algorithm based on the Dijkstra algorithm, and supports dynamic path replanning; The load balancing module monitors the workload of each sorting mechanism in real time and uses the ant colony optimization algorithm to dynamically allocate sorting tasks, keeping the load difference between each device within 15%; The timing coordination module establishes a device collaboration model through a timed Petri net, achieving microsecond-level synchronization control of the timing of multiple device actions, with an accuracy of 0.5ms between adjacent device actions. The fault redundancy scheduling module has a preset three-level fault response mechanism. When a single device fails, it automatically switches to the backup device with a switching time of less than 200ms.

[0008] Optionally, the Faster R-CNN architecture is: the region proposal network (RPN) generates candidate boxes, the ROI pooling layer extracts fixed-size features, and the classifier and regressor output the target category and location.

[0009] An intelligent logistics sorting method, based on the intelligent logistics sorting system, comprises the following steps: Cargo transportation and information collection: Cargo enters the system through the cargo transportation unit. The visual recognition module and RFID recognition module of the intelligent recognition unit collect various characteristic data and destination information of the cargo and transmit them to the central control unit; Sorting instruction generation: After receiving cargo information, the central control unit combines the preset sorting rules and the current sorting task, uses the intelligent sorting algorithm to analyze and process, determines the optimal sorting path and corresponding sorting mechanism for the cargo, generates sorting instructions, and sends them to the sorting execution unit and equipment scheduling unit; Equipment scheduling and sorting execution: The equipment scheduling unit dispatches the cargo conveying unit and the sorting execution unit according to the instructions to ensure that the cargo is transported to the corresponding sorting mechanism according to the predetermined path. The sorting execution unit starts the corresponding sorting mechanism to complete the sorting operation.

[0010] Beneficial effects of the present invention: 1. The present invention achieves comprehensive information collection of goods by integrating a visual recognition module with an RFID recognition module. The visual recognition module uses a multispectral imaging system, combined with an RGB camera and a near-infrared camera, which can simultaneously acquire visible light and near-infrared images. Combined with pre-processing algorithms such as histogram equalization and Gaussian filtering, it effectively enhances image quality. The deep learning recognition engine based on the Faster R-CNN architecture and the ResNet-50 backbone network can accurately identify the outline, barcode, and text information of goods. The three-dimensional reconstruction unit further calculates the precise dimensions of the goods through structured light scanning and the ICP algorithm. The RFID recognition module utilizes a UHF frequency band reader, a phased array antenna array, and a dynamic frame-slot ALOHA algorithm to achieve fast and accurate reading of multiple tags. The two work together to achieve an accuracy rate of over 99.9% in cargo information recognition, greatly reducing sorting errors caused by misreading information.

[0011] 2. The path planning module of the equipment scheduling unit is based on the improved Dijkstra algorithm and can quickly generate the optimal delivery path based on the cargo destination and the real-time status of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a schematic block diagram of an intelligent logistics sorting system provided by the present invention; Figure 2is a schematic block diagram of an intelligent recognition unit according to an embodiment of the present invention; Figure 3 is a schematic block diagram of an implementation of a device scheduling unit according to the present invention; Figure 4 It is a schematic flow chart of the intelligent logistics sorting method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0013] The following will clearly and completely describe the concept and technical effects of the present invention in conjunction with the embodiments so as to fully understand the purpose, characteristics and effects of the present invention.

[0014] An intelligent logistics sorting system, such as Figure 1 As shown, the system comprises a cargo conveying unit, an intelligent identification unit, a central control unit, a sorting execution unit, and an equipment scheduling unit. The cargo conveying unit consists of multiple interconnected conveyor belts equipped with speed regulators and photoelectric sensors to transport cargo to the sorting area. The speed regulator adjusts the conveyor belt speed in real time based on cargo type and sorting requirements, while the photoelectric sensors detect cargo position and transit status. The central control unit receives cargo information from the intelligent identification unit, analyzes and processes the data based on the system's preset sorting rules and the current sorting task, generates sorting instructions, and sends them to the sorting execution unit and equipment scheduling unit. It also monitors the operating status of equipment within the system in real time. The sorting execution unit contains multiple sorting mechanisms that sort cargo according to the sorting instructions from the central control unit. These sorting mechanisms include drive devices and positioning devices. Based on the instructions from the central control unit, the equipment scheduling unit centrally schedules and manages the cargo conveying unit's operating path, the sorting execution unit's operating sequence, and the intelligent identification unit's operating hours.

[0015] like Figure 2As shown, the intelligent recognition unit includes a visual recognition module and an RFID recognition module. The visual recognition module identifies the size, shape, color, and label information of the goods, while the RFID recognition module reads the RFID tag information of the goods via radio frequency. Specifically, the visual recognition module includes a multispectral imaging system, an image preprocessing module, a deep learning recognition engine, a 3D reconstruction unit, and a tag anti-collision processing module. The multispectral imaging system is equipped with an RGB camera and a near-infrared camera to simultaneously capture visible light and near-infrared images of the goods. The near-infrared camera operates in the wavelength range of 700-1000nm, improving the ability to recognize goods of different materials and colors. The image preprocessing module uses histogram equalization and Gaussian filtering to reduce noise and enhance the original image. The specific steps are: original image input → grayscale transformation → histogram correction → enhanced image output. The Gaussian filtering process is: convolution kernel generation → image convolution operation → denoised output. The deep learning recognition engine builds an object detection model based on the Faster R-CNN architecture, combining it with a ResNet-50 backbone network to extract image features. A feature pyramid network performs multi-scale detection, enabling recognition of product outlines, barcodes, and text. The Faster R-CNN architecture consists of a region proposal network (RPN) generating candidate bounding boxes, a region of interest (ROI) pooling layer extracting fixed-size features, and a classifier and regressor outputting object category and location. The 3D reconstruction unit uses a structured light 3D scanner to acquire 3D point cloud data of the product. It then uses an iterative closest point algorithm to align the point clouds and calculate the precise dimensions of the product. The specific steps are: camera capture of the deformed pattern → 3D point cloud generation → ICP algorithm registration → dimension calculation. The tag anti-collision processing module uses the dynamic frame-slot ALOHA algorithm and can simultaneously recognize over 80 tags. The specific steps are: initial frame length setting → random tag slot selection → collision detection → dynamic frame length adjustment → binary tree splitting → tag-by-tag recognition.

[0016] Specifically, the drive device includes a servo motor drive system and a hydraulic drive system. The servo motor drive system uses an encoder to provide real-time feedback on the motor speed and position information to achieve high-precision speed and position control. The hydraulic drive system uses a hydraulic pump to provide stable pressure and accurately adjusts the hydraulic oil flow and pressure through a hydraulic valve group to drive the actuator of the sorting mechanism; the positioning device includes a laser positioning sensor and a grating scale positioning device. The laser positioning sensor accurately measures the position of the goods and the position of the moving parts of the sorting mechanism by emitting a laser beam and receiving reflected light. The grating scale positioning device reads the scale information on the grating scale.

[0017] In one embodiment, the equipment scheduling unit includes a path planning module, a load balancing module, a timing coordination module, and a fault redundancy scheduling module. The path planning module constructs a path topology graph for the logistics sorting system based on a modified Dijkstra algorithm. Physical paths, such as conveyor belts and sorting lanes, are abstracted as edges in the graph, while intersections and sorting ports are abstracted as nodes. Each node and edge is assigned a dynamic weight, such as the degree of conveyor congestion and the busyness of the sorting mechanism. Photoelectric sensors deployed at key locations on the conveyor belts monitor cargo position and flow in real time, updating the path topology graph weights every 50 milliseconds. When a congestion is detected on a conveyor belt section, the system immediately triggers path replanning, calculating an alternative route within 100 milliseconds and sending speed adjustment and steering commands to cargo transport units to ensure efficient detours. Path priorities are set for different cargo types. For example, fresh produce has the highest priority, and the system prioritizes the shortest, uncongested route for them. For general cargo, routes that balance efficiency and energy efficiency are selected while ensuring the passage of priority cargo.

[0018] The load balancing module installs current sensors and pressure sensors on key locations such as the drive motors and conveyor belts of each sorting mechanism (push-plate sorters, slider sorters, etc.) to collect data such as equipment operating current and load weight in real time. Combined with the equipment processing capacity model, the module calculates the current load rate of the equipment every 200ms.

[0019] Application of the Ant Colony Optimization Algorithm: Using machine load rates as inspiration, the algorithm simulates ant foraging behavior and dynamically adjusts sorting task allocation. When the load rate of a cross-belt sorter exceeds 80%, the algorithm automatically reassigns subsequent shipments to a slider sorter with a load rate below 50%, ensuring that the load variance across machines remains within ±15%. Task pre-assignment mechanism: Based on the cargo information (size, weight, destination, etc.) provided by the intelligent identification unit, the processing time of each sorting organization is predicted in advance, and task pre-assignment is completed 1 second before the cargo arrives at the sorting area, reducing equipment waiting time.

[0020] The timing coordination module abstracts operations such as cargo transportation, identification, and sorting into places and transitions in the time Petri net. Each transition corresponds to an equipment action (such as conveyor belt start and stop, sorting mechanism push action) and is assigned precise time parameters (such as the slider sorting machine push action takes 500ms). Microsecond-level synchronization control: Utilizing the IEEE 1588 precision clock protocol for industrial Ethernet, clock synchronization is achieved across all device controllers, with an error controlled within ±1μs. 500ms before a shipment reaches the visual recognition area, the system automatically triggers the camera to take a photo. Once recognition is complete, the sorting mechanism is controlled with microsecond precision to initiate action when the shipment reaches the corresponding sorting port, ensuring zero delay in action. Conflict Detection and Resolution: A conflict detection matrix is ​​established for equipment movement, monitoring the timing of each equipment movement in real time. When two sorting mechanisms are detected as potentially occupying the same conveyor belt area, the system automatically adjusts the movement timing of the subsequent arriving equipment based on task priority and estimated completion time, avoiding equipment collisions and cargo congestion.

[0021] The fault redundancy scheduling module includes: Level 1 response (minor abnormality): When the equipment has minor abnormalities such as overtemperature or motor speed fluctuation, the system will immediately issue an early warning, automatically reduce the equipment's task load by 20%, and start the equipment self-test program to locate the cause of the fault within 5 seconds.

[0022] Secondary response (partial function failure): If the antenna of an RFID identification module fails, the system switches to the backup antenna within 100ms and reallocates the identification task to ensure uninterrupted collection of cargo information.

[0023] Level 3 response (complete equipment failure): When a core sorting mechanism fails completely, the system reconstructs the path topology and task allocation plan within 200ms, diverts all tasks of the faulty equipment to other available equipment, and notifies maintenance personnel through sound and light alarms.

[0024] This embodiment provides an intelligent logistics sorting method, which is based on the intelligent logistics sorting system and includes the following steps: Cargo transportation and information collection: Cargo enters the system through the cargo transportation unit. The visual recognition module and RFID recognition module of the intelligent recognition unit collect various characteristic data and destination information of the cargo and transmit them to the central control unit.

[0025] Sorting instruction generation: After receiving the cargo information, the central control unit combines the preset sorting rules and the current sorting task, uses the intelligent sorting algorithm to analyze and process, determines the optimal sorting path and corresponding sorting mechanism for the cargo, and generates sorting instructions that are sent to the sorting execution unit and the equipment scheduling unit.

[0026] Equipment scheduling and sorting execution: The equipment scheduling unit dispatches the cargo conveying unit and the sorting execution unit according to the instructions to ensure that the cargo is transported to the corresponding sorting mechanism according to the predetermined path. The sorting execution unit starts the corresponding sorting mechanism to complete the sorting operation.

Claims

1. An intelligent logistics sorting system, characterized in that: include: The cargo conveying unit consists of multiple interconnected conveyor belts equipped with speed regulators and photoelectric sensors to transport cargo to the sorting area. The speed regulator adjusts the conveyor belt speed in real time according to the cargo type and sorting requirements, and the photoelectric sensor detects the cargo position and passing status. The intelligent identification unit includes a visual recognition module and an RFID recognition module. The visual recognition module identifies the size, shape, color and label information of the goods, and the RFID recognition module reads the RFID tag information of the goods through wireless radio frequency; The central control unit receives cargo information from the intelligent identification unit, analyzes and processes the data based on the system's preset sorting rules and current sorting tasks, generates sorting instructions, and sends them to the sorting execution unit and equipment scheduling unit. It also monitors the operating status of the equipment in the system in real time. The sorting execution unit includes multiple sorting mechanisms, which sort the goods according to the sorting instructions of the central control unit. The sorting mechanism includes a driving device and a positioning device; The equipment scheduling unit, based on the instructions of the central control unit, uniformly schedules and manages the running path of the cargo conveying unit, the working sequence of the sorting execution unit, and the working time of the intelligent identification unit.

2. The intelligent logistics sorting system according to claim 1, characterized in that: The visual recognition module includes: a multispectral imaging system configured with an RGB camera and a near-infrared camera for simultaneously capturing visible light images and near-infrared images of the goods, wherein the near-infrared camera operates in the wavelength range of 700-1000nm; Image preprocessing module, which uses histogram equalization to reduce noise and enhance the original image; The deep learning recognition engine builds an object detection model based on the Faster R-CNN architecture, combines it with the ResNet-50 backbone network to extract image features, and uses a feature pyramid network to achieve multi-scale detection, identifying cargo outlines, barcodes, and text. The 3D reconstruction unit uses a structured light 3D scanner to obtain 3D point cloud data of the goods, uses an iterative closest point algorithm to perform point cloud registration, and calculates the precise dimensional parameters of the goods; The tag anti-collision processing module is based on the dynamic frame time slot ALOHA algorithm and can identify more than 80 tags at the same time.

3. The intelligent logistics sorting system according to claim 2, characterized in that: The Faster R-CNN architecture is as follows: the region proposal network (RPN) generates candidate boxes, the ROI pooling layer extracts fixed-size features, and the classifier and regressor output the target category and location.

4. The intelligent logistics sorting system according to claim 1, characterized in that: The drive device includes a servo motor drive system and a hydraulic drive system. The servo motor drive system uses an encoder to provide real-time feedback on motor speed and position information to achieve high-precision speed and position control. The hydraulic drive system uses a hydraulic pump to provide stable pressure and accurately adjusts the hydraulic oil flow and pressure through a hydraulic valve group to drive the actuator of the sorting mechanism. The positioning device includes a laser positioning sensor and a grating ruler positioning device. The laser positioning sensor accurately measures the position of the goods and the position of the moving parts of the sorting mechanism by emitting a laser beam and receiving reflected light. The grating ruler positioning device reads the scale information on the grating ruler.

5. The intelligent logistics sorting system according to claim 1, characterized in that: The equipment scheduling unit includes: a route planning module that generates the optimal delivery route based on the cargo destination and current equipment status, and supports dynamic route replanning; The load balancing module monitors the workload of each sorting mechanism in real time and uses the ant colony optimization algorithm to dynamically allocate sorting tasks, keeping the load difference between each device within 15%; The timing coordination module establishes a device collaboration model through a timed Petri net, achieving microsecond-level synchronization control of the timing of multiple device actions, with an accuracy of 0.5ms between adjacent device actions. The fault redundancy scheduling module has a preset three-level fault response mechanism. When a single device fails, it automatically switches to the backup device with a switching time of less than 200ms.

6. An intelligent logistics sorting method, based on the sorting method of the intelligent logistics sorting system according to any one of claims 1 to 5, characterized in that: The following steps are involved: Cargo transportation and information collection: Cargo enters the system through the cargo transportation unit. The visual recognition module and RFID recognition module of the intelligent recognition unit collect various characteristic data and destination information of the cargo and transmit them to the central control unit; Sorting instruction generation: After receiving cargo information, the central control unit combines the preset sorting rules and the current sorting task, uses the intelligent sorting algorithm to analyze and process, determines the optimal sorting path and corresponding sorting mechanism for the cargo, generates sorting instructions, and sends them to the sorting execution unit and equipment scheduling unit; Equipment scheduling and sorting execution: The equipment scheduling unit dispatches the cargo conveying unit and the sorting execution unit according to the instructions to ensure that the cargo is transported to the corresponding sorting mechanism according to the predetermined path. The sorting execution unit starts the corresponding sorting mechanism to complete the sorting operation.

Images